Endocrine Journal 2013, 60 (3), 283-290

The Two main causes of hyperglycemia in type 2 diabetes mellitus are impaired insulin secretion and increased insulin resistance [1, 2]. Evaluation of insulin resistance (or sensitivity) and β-cell function is impor-tant for understanding the disease status and selection of pharmacologic treatment. The gold standard of eval-uation of insulin sensitivity is glucose clamp test [3]. However, the test is limited to research use and is diffi-cult to perform at every medical institution. Although there are also other tests, they are often complex or inadequate [4, 5]. Homeostasis model assessment, first described by Matthews et al., is a method for esti-mating insulin sensitivity [6]. This model is based on

Homeostasis model assessment of insulin resistance for evaluating insulin sensitivity in patients with type 2 diabetes on insulin therapy

Kohei Okita, Hiromi Iwahashi, Junji Kozawa, Yukiyoshi Okauchi, Tohru Funahashi, Akihisa Imagawa and Iichiro Shimomura

Department of Metabolic Medicine, Graduate School of Medicine, Osaka University, Suita 565-0871, Japan

Abstract. Homeostasis model assessment of insulin resistance (HOMA-IR) is a simple and useful method for evaluating insulin sensitivity. But it is difficult to apply to type2 diabetes patients treated with insulin. We have devised a method for measuring HOMA-IR and investigated the validity of HOMA-IR for evaluating insulin sensitivity in patients with type 2 diabetes on insulin therapy. In the first arm of the study, 19 poorly controlled diabetic subjects were treated with insulin and underwent euglycemic clamp study. Then the relationship between insulin resistance index assessed by the clamp test (clamp-IR) and HOMA-IR was investigated in these subjects. Log transformed HOMA-IR correlated with log transformed M/I values derived from the standard euglycemic clamp (r=-0.753, p=0.002). In the second arm of the study, we investigated the relationship between HOMA-IR and various clinical parameters in 156 patients with poorly controlled diabetes after glycemic control. Log transformed HOMA-IR correlated negatively with age (r=-0.292, p=0.0002), HDL-C (r=-0.342, p<0.0001), log transformed serum adiponectin (r=-0.309, p=0.0006) and log transformed KITT (r=-0.264, p=0.0009), and positively with body mass index (r=0.499, p<0.0001), waist circumstance (r=0.461, p<0.0001), visceral fat area (r=0.401, p<0.0001), diastolic blood pressure (r=0.223, p=0.0054), log transformed triglyceride (r=0.497, p<0.0001), urinary CPR (r=0.216, p=0.0099), ΔCPR of glucagon stimulation test (r=0.496, p<0.0001) and log transformed insulinogenic index (r=0.325, p=0.0002). These results suggest that HOMA-IR is a useful test for the evaluation of insulin sensitivity even in patients with type 2 diabetes treated with insulin.

Key words: Homeostasis model assessment of insulin resistance (HOMA-IR), Glucose clamp test, Insulin therapy

the theory of a feedback loop between β cells and the liver [7]. The homeostasis model assessment of insulin resistance (HOMA-IR), calculated from fasting plasma glucose level and immunoreactive insulin (IRI), is a simple method for evaluation of insulin sensitivity and correlates with the results of glucose clamp test in sub-jects with mild diabetes without significant hypergly-cemia [8]. Neverthless it is difficult to apply to patients with poor glycemic control [9], those with severe β cell dysfunction [10] or those treated with insulin.

Chronic hyperglycemia is known to induce insu-lin secretion defect and worsen insulin resistance [11]. This phenomenon, called glucotoxicity, is partly revers-

Submitted Aug. 24, 2012; Accepted Oct. 17, 2012 as EJ12-0320Released online in J-STAGE as advance publication Nov. 10, 2012Correspondence to: Kohei Okita, Department of Metabolic Medicine, Graduate School of Medicine, Osaka University, 2-2-B5 Yamadaoka, Suita 565-0871, Japan. E-mail: oki@endmet.med.osaka-u.ac.jpAbbreviations: HOMA-IR : homeostasis model assessment of

original

©The Japan Endocrine Society

insulin resistance, IRI :immunoreactive insulin, BMI :body mass index, FPG :fasting plasma glucose, BMI:body mass index, eVFA: estimated visceral fat area, CPR:C-reactive protein, ΔCPR: increment of CPR with the glucagon stimulation test, M/I values: insulin sensitivity index estimated with the clamp test, KITT: insulin sensitivity index estimated with the insulin tolerance test, I.I.: insulinogenic index.

284 Okita et al.

tion between HOMA-IR and M/I values derived from the standard euglycemic clamp was investigated.

HOMA-IR was calculated using the following for-mula: HOMA-IR = FPG (mg/dL) × fasting IRI (μU/mL)/405. Before HOMA-IR was calculated, patients were switched to treatment with sulfonylurea (gliben-clamide 1.25 or 2.5 mg) instead of NPH insulin at the night of the day before the measurement to minimize the influence of insulin injected subcutaneously.

The euglycemic-hyperinsulinemic clamp was per-formed according to the method of DeFronzo et al. [3] with a little modification using an artificial pan-creas (model STG-22, Nikkiso, Tokyo, Japan). Briefly, the test consisted of a 120-min euglycemic hyperinsu-linemic clamp period. During the clamp test, subjects received primed-constant infusion of regular insulin (1.45 mU/kg min, Eli Lilly, Indianapolis, IN) and an exogenous glucose infusion to maintain blood glucose levels at 100 mg/dL and to achieve the desired steady-state serum insulin level (100 μU/mL). When the rate of exogenous glucose infusion reached a steady-state level, we evaluated insulin sensitivity as the average glucose infusion rate during the last 30 minutes divided by the average serum insulin level during the last 30 minutes (M/I).

Study 2The study subjects were 156 Japanese with poorly

controlled type 2 diabetes (79 men and 77 women) who had been admitted to Osaka University Hospital for glycemic control between 2001 and 2008. The clin-ical characteristics of the patients are listed in Table 2. Height and waist circumstance were measured in

ible [12, 13]. Glycemic control is required before eval-uation of insulin sensitivity in patients with poor gly-cemic control. In this regard, insulin sensitivity should be evaluated after the control of blood glucose level in diabetic subjects. HOMA-IR can be used for evalua-tion of insulin resistance in patients on diet therapy or sulfonylureas [14] but might be not suitable for those on insulin therapy, because insulin treatment affects serum insulin levels, which in turn influences the feed-back system between the liver and β cells. While it is necessary to evaluate insulin resistance in insulin users, HOMA-IR can only be used to evaluate insulin resis-tance in such patients after minimization of the effect of subcutaneously injected insulin.

In this study, insulin resistance was evaluated with HOMA-IR in patients on short acting insulin with or without sulfonylureas. The aim of this study was to validate HOMA-IR in patients with insulin-induced glycemic control. First, we treated patients with poor glycemic control with insulin. Then, we evaluated the agreement between HOMA-IR and clamp-IR of subjects on insulin therapy (Study 1). After confirm-ing the validity of HOMA-IR in representing insulin resistance, we investigated the relationship between HOMA-IR and various clinical and biological param-eters that are associated with diabetes to determine the clinical usefulness of HOMA-IR (Study 2).

Materials and Methods

Study 1The study subjects were 19 Japanese type 2 diabet-

ics [12 men and 7 women, aged 53.6±14.9 years, body mass index (BMI) 23.3±5.5 kg/m2, hemoglobin A1c (HbA1c) 8.7±1.2 %] who had been admitted to Osaka University Hospital for glycemic control between 2001 and 2006. The clinical characteristics of the patients are summarized in Table 1. On admission, all oral hypoglycemic agents were withdrawn, and all subjects were treated with diet(25-30 kcal/ kg standard body weight / day) and insulin (regular or ultrarapid insu-lin before each meal) for at least 2 weeks until fast-ing plasma glucose (FPG) fell to less than 140 mg/dL. NPH insulin was added before sleep in 10 sub-jects because their fasting plasma glucose was more than 140 mg/dL, though plasma glucose before sleep was less than 140 mg/dL. When FPG decreased to less than 140 mg/dL after treatment, insulin sensitivity was evaluated with HOMA-IR and clamp-IR. The correla-

Table 1 Characteristics of the subjects of Study 1Males/females 19 (12 / 7) Age (years) 53.6 ±14.9 Body weight (kg) 60.0±19.1Body mass index (kg/m2 ) 23.3±5.5 HbA1c (%) 8.7±1.2 Fasting plasma glucose (mg/dL) 120.0±15.1Fasting C-peptide (ng/mL) 1.77±0.81Fasting immunoreactive insulin (μU/mL) 8.2±7.6Insulin dose (U/day) 27.2±27.9HOMA-IR 2.45±2.38

Data were collected after glycemic control, except for HbA1c, and expressed means±SD. HOMA-IR: homeostasis model assessment of insulin resistance

285HOMA-IR in insulin-treated diabetics

at 30 minutes after the 75g glucose load (Δinsulin 0-30 min / ΔPG 0-30 min).

Daily urine samples were collected for measure-ments of urinary CPR. Venous blood sample were collected before breakfast for measurements of LDL-cholesterol, HDL-cholesterol, triglyceride and adi-ponectin. Plasma adiponectin levels were determined with an adiponectin ELISA kit (Otsuka Pharmaceutical Co., Tokushima, Japan), as described previously [17].

The cases with insulin antibody that might have influence on glucose homeostasis were excluded from the studies.

Written informed consent was obtained from all sub-jects, and the study was approved by the ethics com-mittee of Osaka University.

Statistical analysisData are expressed as mean±standard deviation

(SD). Pearson’s correlation coefficient analysis was used to assess the relationship between HOMA-IR and various variables. A p value less than 0.05 was consid-ered significant. All analyses were performed using the Statview 5.5 software (SAS Institute, Cary, NC).

standing position. Visceral fat area was estimated by bioelectrical impedance analysis (BIA), as described previously [15]. On admission, patients were being treated with diet alone (n=29, 18.6%), diet and hypo-glycemic agents (n=103, 66.0%), or diet and insulin (n=24, 15.4%). After admission, oral hypoglycemic agents were withdrawn in all but 9 patients, 24 subjects were treated with diet (25-30 kcal/ kg standard body weight / day) alone, 9 were treated with diet and sul-fonylureas, and 123 with insulin. Only regular or ultr-arapid insulin was used before each meal for at least 2 weeks until FPG decreased to less than 140 mg/dL. When FPG was more than 140 mg/dL while plasma glucose before going to bed was less than 140 mg/dL, NPH insulin was added before sleep.

HOMA-IR was calculated as study1. Then we inves-tigated the relationship between HOMA-IR and vari-ous parameters (age, BMI, waist circumstance, eVFA, systolic blood pressure, diastolic blood pressure, log transformed triglycerides, LDL-cholesterol, HDL-cholesterol, HbA1c, urinary CPR, ΔCPR, log trans-formed insulinogenic index, log transformed serum adiponectin and log transformed KITT).

With regard to antihypertensive and hypolipidemic medications used at admission, 51.1% of subjects were treated with antihypertensive agents and 36.0% of subjects were treated with hypolipidemic agents. These agents were continued until improvement of glycemic control.

Insulin tolerance test was carried out before break-fast after an overnight fast. Patients on NPH insulin were switched to sulfonylurea (glibenclamide 1.25 or 2.5 mg) at the night of the day before the test. Venous blood samples were collected for measurement of plasma glucose before and at 3, 6, 9, 12, 15 minutes after an intravenous bolus injection of regular insu-lin (Novorin R 0.1 U/kg body weight). Fifteen min-utes after insulin injection, the test was terminated by injection of glucose. Insulin sensitivity (KITT) was calculated from the linear slope of the plasma glucose concentration from 3 to 15 minutes, as described pre-viously [16].

The glucagon stimulation test was performed by intravenous infusion of 1 mg glucagon (Novo Nordisk Pharma, Tokyo) after an overnight fast. Blood samples were collected at 0 and 6 min for measurement of CPR. ΔCPR were expressed as increment of CPR. We also calculated the insulinogenic index(I.I.), defined as the ratio of increment in insulin to that in plasma glucose

Table 2 Characteristics of the subjects of Study 2Males/females 156 (79 / 77) Age (years) 60.1±11.5Body mass index (kg/m2) 23.9±4.3Waist circumference (cm) 89.6±12.2 (n=136) Estimated visceral fat area (cm2 ) 107.6±53.1 (n=102) Systolic blood pressure (mmHg) 127.7±17.5Diastolic blood pressure (mmHg) 73.2± 10.8LDL-C (mg/dL) 113.2±26.0 HDL-C (mg/dL) 48.8±14.0 Triglycerides (mg/dL) 102.3±45.7HbA1c (%) 9.4±1.7 Fasting plasma glucose (mg/dL) after treatment 114± 18Fasting immunoreactive insulin (μU/mL) 7.1± 5.1HOMA-IR 2.0±1.3 Urinary C-peptide (μg/day) 65.4±44.6 (n=142) ΔCPR (ng/mL) 2.2±1.2 (n=126) Insulinogenic Index 0.20±0.25 (n=140) adiponectin (μg/mL) 5.4±3.3 KITT (%/min) 1.92±1.22

Data are collected after glycemic control, except for HbA1c, and expressed means±SD. HOMA-IR: homeostasis model assessment of insulin resistance, ΔCPR: increment of C-peptide from the glucagon stimulation test, KITT: K value from insulin tolerance test.

286 Okita et al.

inantly regulated by feedback loop between the liver and β cells [7]. Increased insulin resistance in the liver increases insulin secretion to stabilize hepatic glucose efflux. When the ability of β cells to secrete insulin is appropriate against insulin tolerance, plasma glucose level remains normal. However, defective β cell func-tion results in increased hepatic glucose efflux and con-sequently leads to hyperglycemia. A rise in FPG from 80 to 140 mg/dL results in an increase in fasting plasma insulin, and increases in FPG beyond 140 mg/dL are

Results

Study 1The mean insulin dose used to induce glycemic

control was 27.2±27.9 U/day and FPG improved from 181.1±45.0 to 120.0±15.1 mg/dL. Ten subjects required NPH insulin for glycemic control, and sulfo-nylurea instead of NPH insulin was used at the night of the day before measurement of IRI and to calculate HOMA-IR. After treatment of patients with poor dia-betic control with insulin, fasting IRI was 8.2±7.6 μU/mL and HOMA-IR was 2.45±2.38 (range: 0.77-9.01). M/I value derived from the standard euglycemic clamp test was 0.0464±0.0219 mg/kg/min/μU/mL (range: 0.0067-0.0976).

The correlation between log transformed HOMA-IR and log transformed M/I values derived from the stan-dard euglycemic clamp was significant (r=-0.753, p=0.002, Fig. 1).

Study 2After treatment, the mean fasting plasma glucose

of 156 subjects improved from 178±51 to 114±18 mg/dL. The insulin dose used for glycemic control was 19.1±13.1 U/day. NPH insulin was used in 51 patients for glycemic control, sulfonylurea instead of NPH insu-lin was used at the night of the day before measurement of IRI and to calculate HOMA-IR. After treatment of patients with poor glycemic control, fasting IRI was 7.1±5.1 μU/mL and HOMA-IR was 2.0±1.3. In all of these patients, age (r=-0.292, p=0.0002), HDL-C (r=-0.342, p<0.0001), log transformed KITT (r=-0.264, p=0.0009), log transformed adiponectin (r=-0.309, p=0.0006) correlated negatively with log transformed HOMA-IR after glycemic control. On the other hand, BMI (r=0.499, p<0.0001), waist circumstance (r=0.461, p<0.0001), eVFA (r=0.401, p<0.0001), diastolic blood pressure (r=0.223, p=0.0054), log transformed trig-lyceride (r=0.497, p<0.0001), urinary CPR (r=0.216, p=0.0099), ΔCPR (r=0.496, p<0.0001) and log trans-formed insulinogenic index (r=0.325, p=0.0002) cor-related positively with the log transformed HOMA-IR (Fig. 2). Log transformed HOMA-IR did not corre-late with systolic blood pressure, LDL-cholesterol or HbA1c (Table 3).

Discussion

FPG and serum insulin concentration are predom-

Table 3 Correlation analysis of log transformed HOMA-IR and various clinical parameters

r pAge (years) -0.292 0.0002Body mass index (kg/m2 ) 0.499 <0.0001Waist circumference (cm) 0.461 <0.0001Estimated visceral fat area (cm2 ) 0.401 <0.0001Systolic blood pressure (mmHg) 0.121 0.1338Diastolic blood pressure (mmHg) 0.223 0.0054Log triglyceride (mg/dL) 0.497 <0.0001LDL-C (mg/dL) 0.006 0.9451HDL-C (mg/dL) -0.342 <0.0001HbA1c (%) 0.027 0.41Urinary C-peptide (μg/day) 0.216 0.0099ΔCPR (ng/mL) 0.496 <0.0001Log insulinogenic index 0.325 0.0002Log adiponectin (μg/mL) -0.309 0.0006Log KITT (%/min) -0.264 0.0009

ΔCPR: increment of C-peptide from the glucagon stimulation test, KITT: K value from insulin tolerance test, HOMA-IR: homeostasis model assessment of insulin resistance.

Fig. 1 Study 1. Relation between insulin sensitivity represented by HOMA-IR and that derived from euglycemic hyperinsulinemic clamp (M/I)

287HOMA-IR in insulin-treated diabetics

morning, although the action of NPH insulin may last until the morning. To diminish the effect of exogenous insulin, NPH insulin was substituted with sulfonylurea at the night before the day of estimation of HOMA-IR. Treatment with sulfonylurea is considered to protect against damage of the feedback system between the liver and β cells. Indeed, Emoto et al. demonstrated that log transformed HOMA-IR correlated well with clamp

associated with reduced insulin secretion and increased hepatic glucose output [18].

To evaluate insulin resistance with HOMA-IR, FPG should be less than 140 mg/d and the feedback system between the liver and β cells should be reconstructed. Injection of a high dose of insulin could affect fast-ing IRI and HOMA-IR. Regular or ultrarapid insu-lin injected before supper is almost cleared in the next

Fig. 2 Study 2. Relation between insulin sensitivity measured by HOMA-IR and various clinical parameters

288 Okita et al.

with obesity. These results suggest that insulin resis-tance, expressed by HOMA-IR, is also associated with obesity in poorly controlled type 2 diabetic patients after insulin therapy. Although 51.1% of the patients were being treated with antihypertensive agents and 36.0% of the same subjects were being treated with hypolipidemic agents at study entry, HOMA-IR cor-related with diastolic blood pressure, HDL-C and TG. These results emphasize the validity of HOMA-IR to reflect insulin resistance even after insulin treatment.

Log transformed HOMA-IR also correlated with various clinical parameters associated with insu-lin secretion. Urinary CPR, ΔCPR and insulinogenic index are parameters that express insulin secretion capacity. Increased insulin secretion seems to be also associated with obesity. Insulin can increase adipos-ity since it is a key hormone in adipogenesis. Age is also thought to correlate with insulin secretion capac-ity since insulin secretion ability is known to decrease with age [26]. This phenomenon is attributed in part to decreased β cell sensitivity to glucose-dependent insu-linotropic polypeptide [27] and reduced β2-adrenergic receptor expression [28].

In non-diabetic subjects, increased insulin resistance increases insulin secretion to maintain plasma glucose level within the normal range. Increased insulin secre-tion might lead to increased adiposity, which enhances the likelihood of development of insulin resistance. In this regard, insulin secretion is reported to correlate with insulin sensitivity in a hyperbolic function in unrelated nondiabetic subjects [29]. However, when β cell fails to maintain insulin secretion against insulin resistance, relative insulin deficiency leads to impaired glucose tol-erance or diabetes [1]. Diabetic subjects do not have adequate insulin secretion capacity to keep blood glu-cose within the normal range, but have insulin secretion capacity enough to enhance fat cell growth and body composition. This means that insulin secretion capacity relates to insulin resistance even in type 2 diabetic sub-jects. In this study, we showed that insulin resistance estimated by HOMA-IR correlated with insulin secre-tion ability estimated by urinary CPR, ΔCPR and insu-linogenic index. This means that insulin secretion cor-relates with insulin sensitivity not only in nondiabetic subjects, but also in type 2 diabetic patients.

In diabetic patients with β cell dysfunction, HOMA-IR may not be accurate [10]. In the present study, insulin secretion ability expressed by ΔCPR of glucagon loading test was 2.1±1.0 ng/mL (range: 0.4-

IR in type 2 diabetics treated with sulfonylureas [14]. Insulin treatment may stimulate immunity, and anti-

bodies to insulin may be produced in subjects treated with insulin. Therefore insulin users might have anti-bodies to insulin and these might have influence on glucose homeostasis. In this case, we cannot evaluate insulin sensitivity exactly. Before evaluating insulin sensitivity, we must consider whether insulin antibody is negative or not. The cases with insulin antibody that might have influence on glucose homeostasis should be excluded.

Study 1 showed significant correlation between log transformed HOMA-IR and log transformed M/I derived from the standard euglycemic clamp even in poorly controlled diabetic patients after treated with insulin. HOMA-IR correlated well with log trans-formed M/I in both highly insulin resistant subjects and low insulin resistant subjects. Furthermore, there was no difference in such relationship between patients who did not need and patients who needed NPH insu-lin for glycemic control. These results suggest that HOMA-IR appropriately expresses insulin sensitivity in type 2 diabetic patients under glycemic control with insulin when insulin regimen was optimized to evalu-ate the insulin sensitivity.

Insulin resistance correlates with obesity (especially visceral fat obesity)[19], hypertension [20], dyslipi-demia [21] or hypoadiponectinemia [22, 23]. In Study 2, we have clarified the relationship between log trans-formed HOMA-IR or HOMA-IR and various clinical parameters. The same result was obtained when the subjects were restricted to insulin users. These param-eters except HbA1c were evaluated after glycemic con-trol, because it was presumed that the original state can be evaluated after correction of glucotoxicity.

Log transformed HOMA-IR correlated well with log transformed KITT. KITT is another method used to evaluate insulin sensitivity [16]. KITT is reported to be safe and reproducible method, and the values cor-relate well with M/I values derived from the euglyce-mic hyperinsulinemic clamp test [24, 25]. It should be emphasized that both KITT and HOMA-IR represent insulin sensitivity well even in poorly controlled dia-betics after insulin treatment.

In this study, log transformed HOMA-IR correlated with various clinical parameters associated with obe-sity. BMI, waist circumstance and eVFA are parame-ters of body composition, HDL-C, diastolic blood pres-sure, TG and adiponectin are parameters associated

289HOMA-IR in insulin-treated diabetics

of HOMA-IR for the evaluation of insulin sensitivity in patients with poorly controlled type 2 diabetes after insulin therapy. The results also showed a close correla-tion between log HOMA-IR and log M/I values derived from the standard euglycemic clamp. Furthermore, HOMA-IR correlated with various clinical parameters even in patients with poorly controlled type2 diabetes after glycemic control with insulin. These results sug-gest that HOMA-IR is a reliable and useful parameter for the evaluation of insulin sensitivity in patients with type 2 diabetes treated with insulin. Further examina-tion is expected.

Appendix

We do not have any potential conflicts of interest rel-evant to this article.

4.8) in Study 1, and 2.2 ±1.2 (range: 0.4-5.6) in Study 2. FPG was controlled in all subjects within 140 mg/dL by insulin therapy with or without sulfonylureas. These findings suggest that we can evaluate insulin resistance with HOMA-IR in patients whose ΔCPR of glucagon loading test is more than 0.4 ng/mL and FPG was well controlled without long-acting insulin.

The insulin secretion capacity of Japanese subjects is lower than that of Caucasian subjects [30]. In Japanese subjects, the point of FPG beyond that insulin secretion reduces seems to be lower than that in Caucasian sub-jects. Reduced insulin secretion and increased hepatic glucose output may begin at the point of FPG lower than 140mg/dL. Further examination about the level of FPG on calculating HOMA-IR is expected.

In summary, the present study suggested a method of measuring HOMA-IR and confirmed the validity

References

1. DeFronzo RA, Bonadonna RC, Ferrannini E (1992) Pathogenesis of NIDDM. A balanced overview. Diabetes Care 15:318-368.

2. Ferrannini E (1998) Insulin resistance versus insulin deficiency in non-insulin-dependent diabetes mellitus: problems and prospects. Endocr Rev 19:477-490.

3. DeFronzo RA, Tobin JD, Andres R (1979) Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol 237:E214-223.

4. Monzollio LU, Hamdy O (2003) Evaluation of insulin sensitivity in clinical practice and in research settings. Nutr Rev 61:397-412.

5. Borai A, Livingstone C, Gordon A, Fems A (2007) The biochemical assessment of insulin resistance. Ann Clin Biochem 44:324-342.

6. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Turner RC, et al. (1985) Homeostasis model assess-ment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28:412-419.

7. Turner RC, Holman RR, Matthews D, Hockaday TD, Peto J (1979) Insulin deficiency and insulin resistance interaction in diabetes: estimation of their relative con-tribution by feedback analysis from basal plasma insu-lin and glucose concentrations. Metabolism 28:1086-1096.

8. Bonora E, Targher G, Alberiche M, Bonadonna RC, Saggiani F, et al. (2000) Homeostasis model assess-ment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity: studies in subjects with various degrees of glucose tolerance and insulin

sensitivity. Diabetes Care 23:57-63. 9. Mari A, Pacini G, Murphy E, Ludvik B, Nolan JJ (2001)

A model-based method for assessing insulin sensitiv-ity from the oral glucose tolerance test. Diabetes Care 24:539-548.

10. Kang ES, Yun YS, Park SW, Kim HJ, Ahn CW, Song YD, et al. (2005) Limitation of the validity of the homeostasis model assessment as an index of insulin resistance in Korea. Metabolism 54:206-211.

11. Del Prato S (2009) Role of glucotoxicity and lipotox-icity in the pathophysiology of Type 2 diabetes mel-litus and emerging treatment strategies. Diabet Med 26:1185-1192.

12. Cavaghan MK (2000) Interactions between insulin resistance and insulin secretion in the development of glucose intolerance. J Clin Invest 106:329-333.

13. Poitout V, Robertson RP (2002) Minireview: Secondary beta-cell failure in type 2 diabetes—a convergence of glucotoxicity and lipotoxicity. Endocrinology 143:339-342.

14. Emoto M, Nishizawa Y, Maekawa K, Hiura Y, Morii H, et al. (1999) Homeostasis model assessment as a clini-cal index of insulin resistance in type 2 diabetic patients treated with sulfonylureas. Diabetes Care 22:818-822.

15. Ryo M, Maeda K, Funahashi T, Matsuzawa Y, Shimomura I, et al. (2005) A new simple method for the measurement of visceral fat accumulation by bioelectri-cal impedance. Diabetes Care 28:451-453.

16. Bonora E, Moghetti P, Zancanaro C, Cigolini M, Querena M, et al. (1989) Estimates of in vivo insulin action in man: comparison of insulin tolerance tests with

290 Okita et al.

Clin Nutr 91:258S-261S.24. Akinmokun A, Selby PL, Ramaiya K, Alberti KG

(1992) The short insulin tolerance test for determination of insulin sensitivity: a comparison with the euglycae-mic clamp. Diabet Med 9:432-437.

25. Hirst S, Phillips DI, Vines SK, Clark PM, Hales CN (1993) Reproducibility of the short insulin tolerance test. Diabet Med 10:839-842.

26. Coordt MC, Ruhe RC, McDonald RB (1995) Aging and insulin secretion. Proc Soc Exp Biol Med 209:213-222.

27. Meneilly GS, Ryan AS, Minaker KL, Elahi D (1998) The effect of age and glycemic level on the response of the beta-cell to glucose-dependent insulinotropic polypeptide and peripheral tissue sensitivity to endogenously released insulin. J Clin Endocrinol Metab 83:2925-2932.

28. Santulli G, Lombardi A, Sorriento D, Anastasio A, Del Giudice C, et al. (2012) Age-related impairment in insu-lin release: the essential role of β(2) -adrenergic recep-tor. Diabetes 61:692-701.

29. Mari A, Ahrén B, Pacini G (2005) Assessment of insu-lin secretion in relation to insulin resistance. Curr Opin Clin Nutr Metab Care 8:529-533.

30. Fukushima M, Suzuki H, Seino Y (2004) Insulin secre-tion capacity in the development from normal glucose tolerance to type 2 diabetes. Diabetes Res Clin Pract 66 Suppl 1:S37-S43.

euglycemic and hyperglycemic glucose clamp studies. J Clin Endocr Metab 68: 374-378.

17. Arita Y, Kihara S, Ouchi N, Takahashi M, Maeda K, et al. (1999) Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity. Biochem Biophys Res Commun 257:79-83.

18. DeFronzo RA, Ferrannini E, Simonson DC (1989) Fasting hyperglycemia in non-insulin-dependent dia-betes mellitus: contributions of excessive hepatic glu-cose production and impaired tissue glucose uptake.Metabolism 38:387-395.

19. Yamashita S (1996) Insulin resistance and body fat dis-tribution. Diabetes Care 19:287-291.

20. DeFronzo RA, Ferrannini E (1991) Insulin resistance. A multifaceted syndrome responsible for NIDDM, obe-sity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease. Diabetes Care 14:173-194.

21. Altinova AE, Toruner F, Bukan N, Yasar DG, Akturk M, et al. (2007) Decreased plasma adiponectin is asso-ciated with insulin resistance and HDL cholesterol in overweight subjects. Endocr J 54:221-226.

22. Ryo M, Nakamura T, Kihara S, Kumada M, Shibazaki S, et al. (2004) Adiponectin as a biomarker of the meta-bolic syndrome. Circ J 68:975-981.

23. Ziemke F, Mantzoros CS (2010) Adiponectin in insu-lin resistance: lessons from translational research. Am J